skills.rest
Yuan1z0825
Yuan1z0825COMMUNITY
Explore all skills by this author →
GitHub

nature-image2ppt

Convert slide images, screenshots, and scanned PDFs into object-level editable PowerPoint decks.

38.3k|2.1k|Updated Apr 24, 2026
Search Tags:#powerpoint#OCR#pptx#image conversion#scanned pdf#slide reconstruction#editable deck
One-click install
npx skills add https://github.com/Yuan1z0825/nature-skills --skill nature-image2ppt
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: nature-image2ppt
Source: https://github.com/Yuan1z0825/nature-skills/tree/main/skills/nature-image2ppt
Command: npx skills add https://github.com/Yuan1z0825/nature-skills --skill nature-image2ppt

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pypdfium2, Pillow, numpy, requests, PyYAML, openai, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Slide screenshots, scanned PDFs, and image-only PPTX files lock content inside flat bitmaps, making text, shapes, flowcharts, and arrows impossible to edit. This Skill reconstructs those pages as native, individually editable PowerPoint objects with rendered QA verification.

Core Features & Use Cases

  • Object-Level Reconstruction: Rebuilds text boxes, cards, circle nodes, connectors, arrows, formulas, and measurable curves as native PowerPoint objects rather than flattened images.
  • Semantic-Region Mixed Reconstruction: Decomposes each page into 3-5 semantic regions, keeping measurable elements native while preserving complex visuals as bounded, replaceable image assets with full provenance records.
  • Rendered QA Gates: Validates every page and the final deck with structure, arrow-atomicity, region-decomposition, and rendered-comparison checks backed by itemized evidence files.
  • Use Case: A researcher has a scanned PDF of conference slides and needs an editable PPTX that preserves the original layout, speaker notes, flowcharts, and knowledge graphs for an upcoming presentation.

Quick Start

Ask the agent to convert your slide screenshots or scanned PDF into an editable PowerPoint file while preserving the original layout and speaker notes.

Frequently Asked Questions about nature-image2ppt

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert slide screenshots into an editable PowerPoint?▼

Provide the slide images to the skill, which normalizes them into a run directory, generates OCR text hints, and reconstructs each page as native PowerPoint objects. The final PPTX is assembled from per-page manifests and validated with rendered QA before delivery.

Can a scanned PDF be converted to an editable PPTX?▼

Yes, scanned PDFs are a supported input type alongside images and image-only PPTX files. The skill recovers text, shapes, and page structure, preserving the original layout and rebuilding flowcharts and arrows as editable objects.

Does image to PPT conversion require an internet connection?▼

No, offline mode works with the built-in text geometry detector, though it does not recognize characters. Online OCR via a Baidu PaddleOCR token and online image generation are optional and can be disabled for sensitive material.

What Python version and dependencies are needed?▼

Python 3.10 or later is required with pypdfium2, Pillow, numpy, requests, PyYAML, and openai installed from requirements.txt. Run the doctor command with --json to verify the environment before starting a conversion.

What are the limitations of reconstructing slides from images?▼

Photos, complex illustrations, and dense knowledge graphs may remain as replaceable image assets when native redraw would lose fidelity. Low-resolution inputs, missing fonts, and complex curves also limit reconstruction accuracy and require rendered comparison review.

Can this skill create a new presentation from notes or a paper?▼

No, it only reconstructs existing visual pages and does not author new decks from notes, outlines, or papers. For authoring a new research presentation from paper content, use the nature-paper2ppt skill instead.

Related Skills

Explore More Education & Research Skills→
taracodlabstaracodlabs
779

ocr-and-documents

Extract readable text from PDFs, scanned images, and Word documents using Python libraries.

Official
Intermediate
IeerIeer
1

pptx

Parse PPTX inputs, apply templates, inject content, and export presentations.

Community
Advanced
dralkhdralkh
74

scientific-schematics

Generate publication-quality scientific diagrams from natural-language prompts.

Community
Advanced
LAF-USLAF-US

canva-branded-presentation

Convert briefs or outlines into branded Canva presentations with a slide-by-slide plan.

Official
Intermediate